484 



cavity of the siphon but is closely applied to the wall (see Fig. I). It is 

 the outer or test-bearing surface of the velum that is applied to the 

 wall. For the most part the conditions in the oral and atrial siphons 

 are the same. 



The test-bearing epithelium of the velum appears to be very active. 

 It may, at the junction of the velum with the siphonal wall, grow down 

 into the wall in a series of hollow processes, which of course will con- 

 tain test. A secondary velum may be formed at the base of the siphon 

 at the level of the sphincter muscle. The processes may pass into this 

 secondary velum. I have in no case found them passing into the oral 

 tentacles. 



■TU^tle 



Fig. I. Longitudinal section of wall of atrial siphon in Cnemidocarpa joannae. 



In the atrial siphons of Styelids these processes of the outer epi- 

 thelium of the velum are particularly well developed and they are re- 

 sponsible for the formation of the atrial tentacles. They grow inwards 

 to the inner epithelium of the velum and pushing this epithelium ahead 

 of them, they project into the cavity of the siphon. As a result of this 

 origin, each atrial tentacle consists of a central core of test and 

 around this two epithelial layers united by a thin layer of connec- 

 tive tissue (see Fig. I). The tentacles are always simple (unbran- 

 ched) K 



1 Ritter figures some of them branched in Dendrodoa tuberculata but I have 

 not found them so in that species. 



